Network investing thought experiment
Say 3 companies simultaneously identify a big network-effect opportunity, and none has a particular tactical advantage. If rational and omniscient, each should invest up to its expected present value, or one third of the industry’s estimated PV. The net present value for the entire industry will then be zero, because all the money was spent up front in the battle for the winner-take-all #1 position.
Now move back to the real world, with uncertain, subjective estimates of market size and odds of success. Humans are known to be at their least rational in estimating NPV in low-odds, high-payoff situations (that’s why Lotto tickets sell). So presumably there is a bias for all three entrants to overestimate both the market size and their own odds of victory.
Thus all three are more likely to overinvest than to underinvest, so the industry NPV is negative.
Critically, the above experiment presumed no tactical advantage, and unlimited capital. This shows the incredible importance of tiny tactical advantages in early-stage network effect markets. In addition to being first mover, choosing the deepest-pocketed, fastest-moving VC is tactically useful in signaling to other hopefuls, as early as possible, that the game is over.